Outer Step 8. Inner Step 45. Batch Size 4

config.json

@@ -268,23 +268,23 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5291446,
-    5291450,
-    5291453,
-    5291457,
-    5291461,
-    5291464,
-    5291468,
-    5291471,
-    5291475,
-    5291478
+    5291639,
+    5291642,
+    5291646,
+    5291649,
+    5291652,
+    5291656,
+    5291659,
+    5291663,
+    5291667,
+    5291670
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 39,
+  "inner_step": 45,
   "inner_steps": 0,
   "last_allreduce_block": 5283755,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:75f47639296ae17c8e0e958c235d968833be059a013b19fbb6212956bb9101ea
+oid sha256:8b9a5158131866cd6d256fe2d901c100917df653a66be6aaf87c7f30fa554ec2
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3a55e28c0614fa4b918a99509bd47d9462777ef4245caeab81e2367996395dec
+oid sha256:8bebb49fb079a9dde80b02081ef9b46748780b7a68fccf05752ad8f7fcb94030
 size 4040701744